Analisis Prognostic dengan Metode Hybrid Support Vector Machine dan Extreme Gradient Boost Sebagai Optimalisasi prediksi Remaining Useful Life

Anggraini, Siska Dwi Anggraini (2026) Analisis Prognostic dengan Metode Hybrid Support Vector Machine dan Extreme Gradient Boost Sebagai Optimalisasi prediksi Remaining Useful Life. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5007211010-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Bearing merupakan komponen penting dalam sistem mesin industri yang kegagalannya dapat menyebabkan kerugian besar. Penelitian ini mengembangkan model prediksi Remaining Useful Life (RUL) berbasis kecerdasan buatan dengan mengombinasikan metode SVM dan XGBoost. SVM dimanfaatkan untuk mengklasifikasikan kondisi bearing berdasarkan fitur statistik sinyal getaran seperti
RMS, kurtosis, dan skewness. Pemodelan SVM berguna untuk meningkatkan akurasi dan efisiensi model. Setelah itu, dilakukan pula seleksi fitur dengan metode Recursive Feature Elimination (RFE) serta reduksi dimensi menggunakan Principal Component Analysis (PCA). Selanjutnya, algoritma XGBoost digunakan untuk melakukan prediksi numerik terhadap sisa umur pakai bearing dengan menerapkan proses penalaan hiperparameter dan teknik validasi silang. Secara keseluruhan, pendekatan hybrid ini memberikan solusi yang komprehensif untuk sistem pemeliharaan berbasis data. Model klasifikasi yangdikembangkan berhasil mencapai tingkat akurasi sebesar 99,18% dengan analisis lanjutan sebagai rangkaian dalam prediksi RUL mendapatkan hasil MAE sebesar 0,01 dan RMSE sebesar 0,01.
================================================================================================================================
Bearings are critical components in industrial machinery systems, where their failure can lead to significant operational losses. This study proposes an artificial intelligence – based model for predicting the Remaining Useful Life (RUL) of bearing by integrating Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) methods. SVM is employed to classify bearing conditions based on statistical features extracted from vibration signals, such as RMS, kurtosis, and skewness. To enhance the model’s performance, feature selection is performed using the Recursive Feature Elimination (RFE) method, followed by dimensionality reduction through Principal Component Analysis (PCA). Subsequently, the XGBoost algorithm is employed to perform numerical prediction of the bearing’s remaining service life, incorporating hyperparameter tuning and cross-validation techniques. Overall, this hybrid approach a comprehensive for solution data-driven maintenance systems. The developed classification model achieved a accuracy rate 99,18%, and further analysis as part of the RUL prediction process resulted in a Mean Absolute Error (MAE) of 0.01 and Root Mean Square Error (RMSE) of 0.01.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bearing, Support Vector Machine, XGBoost
Subjects:	T Technology > TJ Mechanical engineering and machinery > TJ174 Maintenance and repair of machinery
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Mechanical Engineering > 21201-(S1) Undergraduate Thesis
Depositing User:	Siska Dwi Anggraini
Date Deposited:	03 Feb 2026 06:14
Last Modified:	03 Feb 2026 06:14
URI:	http://repository.its.ac.id/id/eprint/131998

Actions (login required)

View Item